Historical trends in the deployment of tinctures

 

Prof. G.R. Sampson

 

 

Heraldry achieves visually distinctive displays through a restricted palette of contrasting tinctures.  The theoretically-available palette has enlarged since the middle ages, with the invention of fur variants such as ermines and pean, and of “stains” alongside the traditional colours.  The heraldic literature sometimes seems to suggest that there has been a significant diversification of the range of tinctures encountered in practice, and perhaps a diversification that needs to be reined in.  For instance, John Brooke-Little writes:[1]

 

[Vert and Purpure], not being primary colours, were used but seldom in medieval heraldry, but have increased in popularity with the growth of paper heraldry.  There is also a number of ‘stains’ … which in my opinion should be avoided …

 

In practice, though, my impression has been that the use of different tinctures has changed less over the centuries than one might imagine, and that those significant changes which have occurred have not always lain in the obvious areas.

 

I studied the question by examining the incidence of tinctures[2] in samples of coats from three periods:  the middle ages, the 19th century, and the present day.

 

For the middle ages, I made a random selection of 154 coats from Joseph Foster’s Some Feudal Coats of Arms,[3] avoiding coats based on the arms of England or France (since there are many of those, and for statistical purposes coats in the sample should be independent of one another).  Where quarterly coats appeared to represent unions between families, I considered just the first quarter.  For this and the other samples, any cadency marks were ignored.

 

For the 19th century, I used the Supplement to the 1884 edition of Burke’s General Armory.  This consists largely of what were then recent grants.  There are also older coats which had been omitted from the main text, but Burke’s wording usually makes it easy to identify these; I discarded them, and also discarded quarterly coats (the component quarters of which were likely to be older than the combinations as wholes).  I randomly selected 100 coats from what remained.  It is possible that a few of these were not 19th-century grants, but I feel confident that the great majority, at least, were devised in that period.

 

For the present day, I used all the recent grants I could find in a run of The Heraldry Gazette from June 1994 to Autumn 2000, particularly in its “Contemporary Heraldry” section, together with the new grants blasoned in Peter Gwynn-Jones’s three-part article “Heraldry of new life peers” in The Coat of Arms.[4]  I included only grants by the English or Scotch authorities, so in the Heraldry Gazette listings I omitted Canadian grants.  (This does not reflect any animus against Canada; Canadian heraldry is rightly developing distinctive traditions of its own, whereas I aimed to study the heraldic correlates of the time variable in isolation from any other factors that could confuse the picture.)  This sample comprised 77 coats.  Not all were granted in the 1990s, but all, I believe, post-date the Second World War.

 

Table A shows, for each tincture found in the samples, the number of coats in which the tincture occurs in the respective samples.    Thus the figure 87 for Argent in the Mediaeval column means that 87 of the 154 mediaeval coats sampled include at least one element blasoned Argent, whether field, charge, or both.

 

	Mediaeval	19th C	Present
Argent	87	74	59
Or	85	60	66
Gules	101	44	43
Azure	53	48	38
Sable	34	46	30
Vert	7	9	13
Purpure	1	–	1
Ermine	9	11	4
Ermines	–	–	1
Erminois	–	1	1
Pean	–	–	1
Vair	5	2	–
Murrey	–	–	2
Proper	2	24	11
totals	384	319	270

 

Table A

 

 

 

The differences among the column totals in Table A are due to the fact that the three samples contain different numbers of coats.  If we divide the respective totals by the number of coats in the three samples, we find that the average number of tinctures per coat has increased over time:

 

 

Mediaeval       2.5

19th C             3.2

Present            3.5

 

 

This reflects the familiar fact that newer coats tend to be more complicated, with more charges and hence more possibilities of using different tinctures, than ancient coats.

 

Because the three samples contain different numbers of coats, it would not be very meaningful to study trends in the use of particular tinctures in terms of the raw numbers in Table A.  One way to make the figures commensurable would be to   examine the percentages of all coats which contain different tinctures at different periods.  However, the finding just quoted, that more recent coats tend to include a greater variety of tinctures than older coats, makes this a poor measure – the percentages will tend to rise with time for any tincture.  For instance, Argent is used in 56.5% of the mediaeval coats, 74.0% of the 19th century coats, and 76.6% of the present-day coats:  its popularity seems to have been growing.  But it has not increased in popularity relative to other tinctures.  If, instead of asking what proportion of coats include Argent, we ask what proportion of all instances of a tincture occurring in a coat are accounted for by the tincture Argent (that is, if we divide the figures for Argent in Table A not into the numbers of coats but into the totals in the last row of that Table), the results are:  Mediaeval 22.7%, 19th C 23.2%, Present 21.9% – no real change.

 

Table B shows the percentage use of the tinctures at different periods, calculated in this latter way.

 

 

	Mediaeval	19th C	Present
Argent	22.7	23.2	21.9
Or	22.1	18.8	24.4
Gules	26.3	13.8	15.9
Azure	13.8	15.0	14.1
Sable	8.9	14.4	11.1
Vert	1.8	2.8	4.8
Purpure	0.3	–	0.4
all furs	3.6	4.4	2.6
Murrey	–	–	0.7
Proper	0.5	7.5	4.1

 

Table B

 

 

We see that Brooke-Little was right to suggest that the use of Vert has increased over the centuries.  But it has not increased all that much; it remains far less common than the other colours except Purpure.  Furthermore, the post-mediaeval innovations in terms of furs and stains turn out, statistically, to have made hardly any impact.  The 19th-century and present-day samples do include some instances of the “new” furs, but the figures for all furs taken together at these later periods vary around and close to the mediaeval figure for ermine and vair alone.  My present-day sample contains two coats which use Murrey, but statistically these are barely visible.

 

A larger palette-expanding effect than any of these (including the increased use of Vert) is the change in incidence of charges blasoned Proper.  The mediaeval sample, unexpectedly, does include two coats in which some charge is blasoned Proper; but proportionately this has increased fifteen times by the 19th century, and the present-day figure, though much lower, is still over half the 19th-century figure.  The high 19th century figure doubtless reflects the taste for naturalistic, “landscape” heraldry during the eighteenth and nineteenth centuries; postwar heraldic taste has tended to revert to the mediaeval concept of bold stylized outlines and colours, but once the possibility of blasoning unusual charges Proper had been admitted, it was too convenient to abandon altogether.[5]

 

There are other discrepancies between columns in Table B which may come as more of a surprise.  Most notably, the figure for Gules is far higher in the middle ages than in either later period.  This change is numerically larger than any of those discussed above.

 

Of course, one could expect to find some fluctuations in the incidence of particular tinctures simply as a matter of chance.  Readers may suspect that my samples are too small to draw reliable conclusions about change in use of standard tinctures.  But the science of statistics offers methods for checking such issues.  I reduced the figures of Table A to a 2 x 2 contingency table, counting instances of Gules v. instances of all other tinctures in the mediaeval sample v. the other two samples combined.  This gave a chi-squared statistic of 19.8, p < < 0.001.  Readers unversed in statistical concepts will probably not want a lengthy explanation here, but the upshot is that the finding that mediaeval heraldry had a special preference for Gules is extremely robust.  The chance that it is a mere statistical fluke of the particular samples studied is negligible.

 

Table B also shows fluctuations in use of Sable which are smaller than in the case of Gules but still noteworthy.  What lies behind these becomes clearer if one distinguishes the use of a tincture as field colour from other occurrences of that tincture in a coat.  I made separate counts in the samples for tinctures used as field, and tinctures which occur in a coat but not as field.  I drew this distinction in a mechanical way, counting the first tincture mentioned in a blason as field and all other tinctures as non-field; thus, for a coat party per fess, the tincture of the upper half was counted as “field”, that of the lower half, as well as that of any charge different in tincture from the upper half of the field, as “non-field”.  Obviously, in a case such as a chequy coat, the distinction drawn in this way is fairly meaningless in practice, but for most coats the tincture counted as “field” will be visually more dominant than others.

 

Table C shows the percentages of the three samples which use the respective tinctures as field.

 

 

	Mediaeval	19th C	Present
Argent	26.6	31.0	23.4
Or	23.4	16.0	15.6
total metals	50.0	47.0	39.0
Gules	24.7	13.0	23.4
Azure	16.2	20.0	19.5
Sable	4.5	14.0	6.5
Vert	1.3	2.0	5.2
Purpure	–	–	–
total colours	46.8	49.0	54.5
all furs	3.2	4.0	5.2
Murrey	–	–	1.3

 

Table C

 

 

A 2 x 2 contingency table of the figures for Sable fields v. non-Sable fields in the 19th century v. the other two periods combined gives X² = 7.44, p < 0.01; again a statistically-robust result.  Systematically, the use of Sable fields was greater in the 19th century than at the earlier or later periods; conversely we see that the use of Gules fields was lower in the 19th century than either earlier or later.

 

Red is the most vivid colour (the hue for which the human eye can detect the greatest degree of “saturation”);[6] black is obviously the gloomiest colour.  We know that the Victorians favoured sub-fusc tones in dress, interior decoration, etc.  It is interesting to find this preference reflected so directly in their heraldry.

 

Table C also suggests a historical development over the three periods whereby (leaving aside the minor contribution of furs and stain) fields of metal have become less frequent and coloured fields more frequent.  One cannot be dogmatic about matters of taste, but I think many people interested in graphic design would agree that small bright motifs on a dark background tend to create an effect of greater elegance and authority than dark motifs on a light background.  (Think for instance of gold lettering and colophon on a leather book-binding, or white-painted numeral and brass “furniture” on the black door of 10 Downing Street.)  Possibly, this aesthetic has been increasingly asserting itself against a mediaeval tradition in which shields really were made of metal so that it was natural to make them look like silver or gold.  However, in this case the differences among the numbers did not achieve statistical significance by the test I applied, so this apparent development could be an accident of sampling.

 

If it is a genuine finding, it is accounted for wholly by Or; the Argent figures fluctuate with no steady trend.  One might comment that, since paper is white, exchanging real-life armour for paper heraldry made Argent fields no less natural than they had been before, whereas it replaced the noble associations of gold with the cowardly and unattractive associations of yellow.

 

For completeness, Table D displays figures for the incidence of “non-field” tinctures, calculated as percentages of tincture/coat pairings on the same basis as Table B.

 

 

	Mediaeval	19th C	Present
Argent	20.0	19.6	21.2
Or	21.3	20.1	28.0
total metals	41.3	39.7	49.2
Gules	27.4	14.2	13.0
Azure	12.2	12.8	11.9
Sable	11.7	14.6	13.0
Vert	2.2	3.2	4.7
Purpure	0.4	–	0.5
total colours	53.9	44.7	43.0
all furs	3.9	4.6	1.6
Murrey	–	–	0.5
Proper	0.9	11.0	5.7

 

Table D

 

 

The special disfavouring of Sable as a field colour in Mediaeval and Present periods does not extend to Sable as a non-field colour, where it does not “set the tone” of the coat as a whole.  The only other point of note is that Gules has not regained at the present day the pre-eminence as a non-field tincture which it lost between middle ages and 19th century.  We moderns, it seems, while we have given up the Victorian taste for positively gloomy coats, have not returned to the mediaeval preference for the brightest possible combinations.  The present-day figures show the various tinctures being used for non-field purposes in relatively equal proportions, perhaps about as equal as is possible given the rules about metal on colour and vice versa (which imply that the two metals must always be individually more frequent than the five colours), and given that Vert and Purpure retain their traditional low frequency.

 

The relative rarity of Vert in particular calls for further comment.  Brooke-Little, in the passage quoted earlier, explained it in terms of green not being a “primary colour”, and with respect to paint-mixing this is true.[7]  Perhaps it was felt that a colour obtained by mixing pigments was less suitable for a gentleman’s arms than a “pure” colour.  But in other ways green seems very suitable for heraldry.  We have seen that the mediaevals favoured bright colours, and green is one of the brightest; the eye can detect greater saturation in the green area of the spectrum than in the blue area.[8]  Those of the world’s languages which make do with few words for colours normally have a word for green but no word for blue, rather than the other way round.[9]  It is not clear why technicalities about mixing of pigments should have outweighed this basic visual virtue.

 

My very tentative suggestion would be that the explanation lies in the background colour-scheme of mediaeval life.  Modern urban environments consist largely of manufactured objects and materials which are made in any colours the consumer desires, and modern technology lets us keep things clean.  In the middle ages, part of the attraction of heraldry surely was that it was difficult to create and maintain bright colours, so they stood out when they appeared.  Much of the visual environment must have been filled with the browns of wood and mud, and the greens of grass and foliage.  Brown had no heraldic virtue, since it was drab as well as common, and it was ignored.  Green had the virtue of brightness, but the fact that so much of Nature is green perhaps made it seem less suitable for distinctive graphic displays than other hues; so it played a minor role in heraldry, and continues to do so.

 

To summarize the main findings of the preceding pages:

 

•             the average number of tinctures per coat has risen from 2.5 in the middle ages to 3.5 today;

•             the most significant widening of the heraldic palette has been the introduction of elements blasoned Proper, which were very frequent in the 19th century though they are rather less frequent today;

•             in the 19th century, the field was much less often Gules and much more often Sable than was the case either earlier or later;

•             as a non-field tincture, Gules was specially favoured in the middle ages but later became just one colour among others;

•             there may have been a trend over the centuries to use metals less often and colours more often as field.